The present teaching relates to determining user interests based on a correspondence graph. In one example, correspondences among a plurality of users are obtained. Connections among at least some of the plurality of users are identified based on the correspondences. A graph is constructed based on the identified connections. Each of the identified connections has a strength determined based on correspondences associated with the identified connection. One or more seed users are selected from the plurality of users. Each of the one or more seed users has at least one known interest. An interest is determined for each of the at least some of the plurality of users based on the graph and the known interests of the one or more seed users.
Legal claims defining the scope of protection, as filed with the USPTO.
1. A method, implemented on a computing device having at least one processor, storage, and a communication platform capable of connecting to a network for determining user interest, comprising: obtaining correspondences among a plurality of users; identifying connections between at least some of the plurality of users based on the correspondences; constructing a graph based on the identified connections, wherein each of the identified connections has a strength determined based on one or more topics determined from content included within one or more of the correspondences associated with that identified connection, wherein the identified connections do not include connections associated with only unidirectional correspondences; selecting one or more seed users from the plurality of users based on at least one implied interest derived from the content included within at least some of the one or more of the correspondences associated with the one or more seed users, wherein the at least one implied interest comprises at least one of the one or more topics, wherein the at least one implied interest is classified as a known interest in response to a determination that a confidence level of the at least one implied interest exceeds a confidence threshold, and wherein each of the one or more seed users has at least one known interest; and determining an interest to be propagated to each of the at least some of the plurality of users based on the graph and the at least one known interest of the one or more seed users.
2. The method of claim 1 , wherein: the correspondences comprise at least one of: email, instant message, short message service (SMS) based communication, or multimedia message service (MMS) based communication; and the interest to be propagated to each of the at least some of the plurality of users comprises the implied interest that is classified as the known interest.
3. The method of claim 1 , wherein the strength for each of the identified connections is further determined by: extracting metadata from the one or more of the correspondences associated with the identified connection; identifying parameters of the one or more of the correspondences from the extracted metadata; and determining the strength based on the identified parameters, wherein the identified parameters comprise at least one of frequency, recency, or direction of the identified connection.
4. The method of claim 1 , wherein the strength for each of the identified connections is further determined by: extracting the content included within the one or more of the correspondences associated with the identified connection; and determining the one or more topics of the content.
5. The method of claim 1 , wherein: the one or more seed users are further selected based on one or more declared user interests of each of the one or more seed users; and the at least one known interest of each of the one or more seed users comprises (i) the one or more declared interests of each of the one or more seed users and (ii) the at least one implied interest of each of the one or more seed users that is classified as the known interest.
6. The method of claim 1 , further comprising: propagating the at least one known interest of each of the one or more seed users to each neighboring user of each of the one or more seed users on the graph, wherein the at least one known interest comprises the at least one implied interest that is classified as the known interest; and adjusting levels of the at least one known interest at each neighboring user of each of the one or more seed users based on the strength of the connection between a corresponding neighboring user and a corresponding seed user.
7. The method of claim 1 , wherein the correspondences comprise direct communications or indirect communications.
8. The method of claim 1 , wherein the strength for each of the identified connections is amplified or attenuated based on a set of weights associated with a corresponding identified connection, and a level of the interest for each neighboring connection of each seed user is adjusted based on the strength between the neighboring connection and the seed user.
9. A system for determining user interest, comprising: a user correspondence analyzing module configured to obtain correspondences among a plurality of users and identify connections between at least some of the plurality of users based on the correspondences; a user correspondence graph building module configured to construct a graph based on the identified connections, wherein each of the identified connections has a strength determined based on one or more topics determined from content included within one or more of the correspondences associated with that identified connection, wherein the identified connections do not include connections associated with only unidirectional correspondences; a seed user determining module configured to select one or more seed users from the plurality of users based on at least one implied interest derived from the content included within at least some of the one or more of the correspondences associated with the one or more seed users, wherein: the at least one implied interest comprises at least one of the one or more topics, the at least one implied interest is classified as a known interest in response to a determination that a confidence level of the at least one implied interest exceeds a confidence threshold, and each of the one or more seed users has at least one known interest; and a user interest inference module configured to determine an interest to be propagated to each of the at least some of the plurality of users based on the graph and the at least one known interest of the one or more seed users.
10. The system of claim 9 , wherein: the correspondences comprise one or more of email, instant message, short message service (SMS) based communication, or multimedia message service (MMS) based communication; and the interest to be propagated to each of the at least some of the plurality of users comprises the implied interest that is classified as the known interest.
11. The system of claim 9 , wherein the strength for each of the identified connections is further determined by: a metadata extracting unit configured to extract metadata from the one or more of the correspondences associated with the identified connection; a parameter determining unit configured to identify parameters of the one or more of the correspondences from the extracted metadata; and a connection strength determining unit configured to determine the strength based on the identified parameters, wherein the identified parameters comprise at least one of frequency, recency, or direction of the identified connection.
12. The system of claim 9 , wherein the user correspondence analyzing module comprises: a content extracting unit configured to extract the content included within the one or more of the correspondences associated with the identified connection; and a topic identifying unit configure to determine the one or more topics of the content.
13. The system of claim 9 , wherein: the one or more seed users are further selected based on one or more declared user interests each of the one or more seed users; and the at least one known interest of each of the one or more seed users comprises (i) the one or more declared interests of each of the one or more seed users and (ii) the at least one implied interest of each of the one or more seed users that is classified as the known interest.
14. The system of claim 9 , wherein the user interest inference module comprises: a user interest propagating unit configured to propagate the at least one known interest of each of the one or more seed users each neighboring user of each of the one or more seed users on the graph, wherein the at least one known interest comprises the at least one implied interest that is classified as the known interest; and a user interest aggregating unit configured to adjust levels of the at least one known interest at each neighboring user of each of the one or more seed users based on the strength of the connection between a corresponding neighboring user and a corresponding seed user.
15. A machine-readable, non-transitory and tangible medium having data recorded thereon for determining user interest, the medium, when read by the machine, causes the machine to perform the following: obtaining correspondences among a plurality of users; identifying connections between at least some of the plurality of users based on the correspondences; constructing a graph based on the identified connections, wherein each of the identified connections has a strength determined based on one or more topics determined from content included within one or more of the correspondences associated with that identified connection, wherein the identified connections do not include connections associated with only unidirectional correspondences; selecting one or more seed users from the plurality of users based on at least one implied interest derived from the content included within at least some of the one or more of the correspondences associated with the one or more seed users, wherein the at least one implied interest comprises at least one of the one or more topics, wherein the at least one implied interest is classified as a known interest in response to a determination that a confidence level of the at least one implied interest exceeds a confidence threshold, and wherein each of the one or more seed users has at least one known interest; and determining an interest to be propagated to each of the at least some of the plurality of users based on the graph and the at least one known interest of the one or more seed users.
16. The medium of claim 15 , wherein the strength for each of the identified connections is further determined by: extracting metadata from the one or more of the correspondences associated with the identified connection; identifying parameters of the one or more of the correspondences from the extracted metadata; and determining the strength based on the identified parameters, wherein the identified parameters comprise at least one of frequency, recency, or direction of the identified connection.
17. The medium of claim 15 , wherein the medium, when read by the machine further causes the machine to perform the following: propagating the at least one known interest of each of the one or more seed users to-each neighboring user of each of the one or more seed users on the graph, wherein the at least one known interest comprises the at least one implied interest that is classified as the known interest; and adjusting levels of the at least one known interest at each neighboring user of each of the one or more seed users based on the strength of the connection between a corresponding neighboring user and a corresponding seed user.
Cooperative Patent Classification codes for this invention. Click any code to explore related patents in that topic.
December 17, 2014
February 11, 2020
Browse 5M+ US patents with plain-English claim translations and AI-generated analysis.